SongKong Jaikoz

SongKong and Jaikoz Music Tagger Community Forum

Acoustic ID retrieval SLOW performance with Jaikoz 4.3.1 / OS X 10.7.1

#1

Paul,

I am trying to successfully tag through a music library of about 15K items.

In the past I have been able to tag libraries of 4K to 5K items using Jaikoz without much of a hassle. But now for some reason the process seems to be running way too slow.

I recently updated to OS X Lion (10.7.1) and Jaikoz 4.3.1 NGS.

The Music Library is stored in a Synology NAS Server connected through Gigabit Ethernet to my iMAC.

This is my workflow for tagging large libraries:

1.- Add the Folder for the new Library to Jaikoz. (Done … took a while but got there for the 15K files)

2.- Filter for Files with NON Acoustic ID.

3.- On the Edit window Select All and Remote Correct -> Retrieve Acoustic IDs (STUCK HERE !!!).

4.- Remote Correct -> Autocorrect Metadata from MusicBrainz. SAVE Changes

4.1- Local Correct -> Delete Duplicates. SAVE Changes

5.- Remote Correct -> Autocorrect Metadata from Discogs. SAVE Changes

6.- Remote Correct -> Correct Lyrics. SAVE Changes

7.- File and Folder Correct -> Correct Filenames from Tags. SAVE Changes

8.- File and Folder Correct -> Correct Sub Folders from Tags. SAVE Changes

For the 15K files scenario I started retrieving Acoustic IDs about 4 hrs ago and only got through around 400 files … I don�t recall this being so slow in the past … with previous libraries of 4K to 5K files.

Am I looking forward at a weeklong process just to retrieve the Acoustic IDs ?

Neither the iMAC nor the NAS Server seem stressed at all in CPU or memory.

The only idea that comes to mind is that the OS must be paging virtual memory … but don’t know how to test or correct this problem for Jaikoz due to the Java architecture.

Are there any Jaikoz or OS settings I can change to improve the performance of the acoustic ID retrieve ?

Are there any test you suggest to perform to further understand the issue ?

Does the workflow I use make sense to you or should I go about tagging a large library in a different way ?

Any other suggestions ?

Please feel free to link to other threads if the answers are there (I did not find any).

Thanks for your help.

#2

(Ive edited your title because the length was preventing me replying due to bug in forum software)

Retrieve Acoustic Ids should now be MUCH quicker because it done locally, but you are loading 15,000 files which is alot for Jaikoz with the default memory settings.

The one preference that doesnt get retained when you upgrade Jaikoz is memory allocated, so I wonder if yo had modified it previously and now lost that modification.

Try increasing that and also send me your support files (Advanced/Create Support Files)

#3

Moved the new library to a local USB HD, turned off antivirus inspector for that location and for the Jaikoz app and now it works fine. So it doesn’t sound like a memory issue.

But still doesnt work well with the NAS. Don’t know why. Any other ideas ?

Will send you my support files for an autocorrect run of the NAS Library. Perhaps of only a few files, otherwise it is never going to finish.

Now I am getting an eventual cannot connect to the MusicBrainz server Error during Autocorrect … I Just repeated the Autocorrect from MusicBrainz action or should I do something else ?

… learned quite a bit about matching and tagging in the process …

Also turned off iTunes Update because I will still scrub these files quite a bit before I move them to the iTunes Library … but that is sitting in the NAS … so I am still not sure how Im going to move them into iTunes and keep the library there cleaned up considering the performance issue with the NAS … any suggestion is welcomed.

Can you point me to any forum threads or documentation about the recommended approaches to tag a very large song Library that is to be used with iTunes ?

Perhaps addresses points like:

  • Recommended strategies for filenames/subfolders

  • Recommended strategy for tagging large quantities of files … what is more efficient ?

  • Recommended Autocorrector Sequences

Hopefully there is something out there so I don�t have to learn everything by trial an error.

Just point me in the right direction … I will do the studying.

Thanks.

#4

I still think you need to increase the memory allocated.

#5

I did.

Doubled the memory as per the instructions.

NAS is still a NO GO … tried with the additional memory and got the same result.

But the memory change did improve processing from the USB HD.

Problem now is that tagging runs are not finishing properly for some file sets.

In the console you can see MusicBrainz Server Connection problems as well as non valid return codes such as null pointers and others.

Eventually when a “worker” reaches 9 re-connection attempts the Remote Correct is cut short with “Problem Connecting to MuscBrainz Server” message.

It seems there are particular File sets that are harder than others in this respect. If I then run Remote Correct for Discogs it is able to match a considerable number of files that would not match for MusicBrainz, and I never get any type of connection error.

What level of logging should I setup in the init file so it is easier to understand what the problems might be while polling MusicBrainz ?

Thanks,

#6

HI, if you are getting errors the first step is for you to send me your support files so I can take a look at them (Advanced/Create Support Files)

#7

I just sent you the support files

#8

Where did you send it I don’t seem to have received it.

#9

Those logs are for Jaikoz using the USB HD as the source for the files

#10

Yeah but where are they, I don’t seem to have them.

#11

If we can get the tagging to work well from the USB HD I will send you a log using the NAS as source … so far due to the MusicBrainz “connection” errors that cancel the process it seem Jaikoz can only get to taging less than 50 % of the files.

Thanks again,

#12

sent to support@jthink.net

I guess you want me to upload them as well ?

#13

Tried Uploading here and I get an Apache Error … zip file is 14.2 Mb

#14

[quote=tupublico]sent to support@jthink.net

I guess you want me to upload them as well ?[/quote]

No I have them, there was delay between you saying you’d sent them and me actually get them.

#15

[quote=tupublico]If we can get the tagging to work well from the USB HD I will send you a log using the NAS as source … so far due to the MusicBrainz “connection” errors that cancel the process it seem Jaikoz can only get to taging less than 50 % of the files.

Thanks again,[/quote]

Musicbrainz has been experiencing a much higher load over the last few weeks ( I only found this out yesterday) and they are making changes to combat this so I expect this will resolve itself with no chnage required to Jaikoz.

#16

On Windows 7 and just spent two days waiting for 2600 MP3s to match and then when I wasn’t watching Windows rebooted the computer to apply updates and now I’m starting over. It is almost the exact situation as the OP so that is why I thought I should mention it here. I’m going to keep searching for some performance enhancing threads but if you’ve made any progress on this please post updates.